Technology holds promise though debt pricing algorithm works well without it 

By Brett Boehm 

Equipment finance businesses are in the early stages of using artificial intelligence (AI) tools to drive improvements in credit and collections. But is AI saving them time? 

That question was top of mind at the Equipment Leasing and Finance Association (ELFA) Credit & Collections Management Conference this year. A variety of important topics were covered in depth at the event, including credit scoring as part of my speaker session, yet discussions about AI and companies introducing AI in their businesses seemed to dominate many conversations. 

My sense is that most companies represented there were still feeling their way through AI developments and trying to figure out how, and when, to adopt the technology in their operations. Some were using AI in a basic way to draft letters or emails, but not seriously in their businesses. That’s what I’m seeing personally, as well, when working directly with equipment finance teams and other lenders as their commercial debt buyer.  

The way most companies are tiptoeing into it, AI is like having a personal assistant whose work you must regularly review for accuracy. Helpful, but not a time saver. Once more companies are confident that AI makes fewer mistakes than humans do, and saves them time, it could be a game changer. 

That’s not to say AI isn’t already gaining some traction in equipment finance. The conference provided an interesting preview of ELFA survey data speaking to this, and I recently read about an idea for an AI initiative targeted at improving collections. But first let me tell you about an industry algorithm that’s working well without AI.  

Accurate debt pricing without AI 

Our company, TBF, is the leading buyer of non-performing equipment leases, commercial bank loans, online small business loans, merchant cash advances and commercial credit card accounts in the United States. When we provide quotes to customers, we must understand the true market value of their distressed accounts and make the pricing competitive. 

To that end, our CFO Adam Boehm developed a data-driven, true market pricing algorithm—MarkTrue—that we use internally for producing customer quotes. The algorithm leverages 28 years of proprietary data on all types of commercial products including equipment leases and loans, in all types of economic climates. It grades customer accounts and compares them to historical data to determine what the ultimate recovery amount will be. This ensures each purchase offer reflects the accounts’ true market value.  

I don’t see a need to leverage AI tools for our algorithm, at least not yet, because it is already efficient and highly accurate. 

AI’s potential for better collections  

At the same time, there is certainly the promise of AI delivering future improvements in credit and collections. One session of the ELFA Credit & Collections Management Conference offered a preview of the 2025 Credit Managers Survey results, which show that momentum for AI is slowly building. Responding to a question about their use of AI for credit underwriting or credit portfolio management, the percentage of survey participants who said they were either exploring, planning or implementing AI has grown since 2024. We look forward to additional insights when the association releases their data publicly. 

Also of interest is a recent Monitor article by Valerie L. Gerard of The Alta Group and Denis Stypulkoski of Reimagine Advisors, who have proposed an AI initiative targeting collections in the equipment finance industry.  

“How do you think big in a generative AI sense, so that the industry moves well beyond incremental process improvements? One approach, discussed here, is to start in collections, by more fully automating early-stage interactions, where conversational natural language models and bots could free up employees to partner with their collections copilot to handle the most challenging late-stage collections problems,” they wrote. “While this is speculative at this point, and might look highly improbable to some, if we don’t start somewhere, how will we ever get to the next level?”